New Delhi: In a quiet sign of just how big tech companies are investing in their AI systems, Microsoft has quietly become one of the biggest purchasers of AI from Meta.
Meta is dispending hundreds of millions of dollars per year to obtain AI models via Microsoft’s Azure cloud computing platform, according to sources close to the situation. But Meta is also consuming a ton of computing power, using trillions of AI tokens per week on the platform.
Why is Meta using Microsoft’s AI services?
One of the largest firms working on their own AI models is Meta. It has made big investments in artificial intelligence data centres, chips and computing infrastructure. But constructing everything internally is very costly and time-consuming.
Rather, Meta is also leveraging AI models created by other firms. Meta is able to leverage various AI models based on their performance, availability and cost through Microsoft’s Azure platform and its Foundry marketplace.
According to the report, Meta developers have used OpenAI technology to evaluate and compare outcomes of their own AI models using Microsoft Foundry. That will provide Meta with an additional platform for testing the capabilities of its technology against some of the top AI systems on the market.
Microsoft Foundry is also a significant addition to Microsoft’s AI enterprise. As of July, around 100,000 customers used the marketplace, which provides access to models from various artificial intelligence providers.
Meta is not depending only on Microsoft
Meta is not abandoning its own AI efforts due to its partnership with Microsoft. Meta’s partnership with Microsoft doesn’t mean it’s steering clear of its own AI development.
Meta is investing billions of dollars in its AI infrastructure and developing its own models. The company also leases access to external AI models as appropriate to its development efforts.
Additionally, Meta is under development of its own AI API business that would potentially present a threat to Microsoft Foundry when customers have the ability to buy and use different AI models from the Meta platform. The move would put a dent in Meta’s reliance on services from Microsoft and other cloud vendors, if it succeeds.
A big opportunity for Microsoft
For Microsoft, it is significant that Meta is spending on its AI capabilities, which indicates that the demand for AI extends beyond the realm of traditional business operations. Technology firms such as ByteDance, Adobe, Perplexity and Sierra are among the largest AI adopters among Microsoft’s customers.
This can pose a problem for Microsoft and the AI industry, however. If revenue from AI is going to be concentrated into a few tech firms, the expansion perhaps isn’t as wide as anticipated.
In the short-term, Meta’s reliance on Microsoft’s AI services reveals that even corporations developing their own AI solutions require significant computing power and access to various models. Microsoft is capitalising on that demand by being an important infrastructure player in the race to AI.
